ES2365176T3 - METHODS TO VALUE THE RISK OF HEART INTERVENTIONS BASED ON GEN GDF-15. - Google Patents

Abstract

Method to find out if a subject affected by an acute coronary syndrome needs a cardiac intervention, which consists in a) determining the amount of the growth differentiation factor (GDF-15) in a sample of blood, serum or plasma of a subject in need of a cardiac intervention and b) comparing the amount of GDF-15 found in stage a) with a reference amount of 1200 pg / ml, so that an amount of GDF-15 greater than 1200 pg / ml indicates that a subject is susceptible to cardiac therapy and said cardiac intervention is chosen from the group consisting of percutaneous coronary angioplasty, percutaneous transluminal coronary balloon angioplasty, laser angioplasty and coronary stent implantation or bypass implantation.

Description

The present invention relates to a method of finding out if a subject is prepared for cardiac intervention, based on the determination of growth differentiation factor-15 (GDF-15) in a sample of a subject in need of cardiac intervention.

One of the aspirations of modern medicine is to provide personalized or individualized treatments, that is, that take into account the needs or risks of each patient. Customized or individualized treatment regimens should also be taken into account for emergency measures. Specifically, in the case of acute cardiovascular events, a certain treatment regimen must be chosen, usually in a short period of time. Cardiovascular complications, particularly heart disease, are the leading cause of morbidity and mortality in the Western Hemisphere. Cardiovascular complications may remain asymptomatic for long periods of time. However, they can have serious consequences when an acute cardiovascular episode occurs, such as myocardial infarction, because of cardiovascular complications.

Conventional techniques for diagnosing cardiovascular complications include electrocardiographic and echocardiographic measurements, symptom analysis and previous patient history, such as chest pain and analysis of some clinical parameters. These conventional techniques have been reinforced lately with the analysis of biomarkers and in particular the levels of cardiac troponins in blood samples from emergency patients. On the other hand, natriuretic peptides are also described as suitable biomarkers to diagnose cardiovascular complications. Even more recently it has been suggested that GDF15 is also an indicator of cardiovascular complications (US 2003/0232385; Kempf 2006, Circ Res 98: 351360). Growth differentiation factor-15 (GDF-15) is a member of the cytokine superfamily that includes transforming growth factors-β. GDF-15 was initially identified as a type 1 macrophage inhibitor cytokine (MIC-1) and subsequently also referred to as placental transforming growth factor-β (Bootcov 1997, Proc Natl Acad Sci 94: 11514-11519; Tan 2000, Proc Natl Acad Sci 97: 109-114). It has recently been shown that cultured cardiomyocytes express and secrete GDF-15 through signaling pathways dependent on nitric oxide and nitrosing stress when subjected to simulated ischemia and reperfusion. Likewise, using a model of ischemia lesion and myocardial reperfusion in mice, it has been observed that GDF-15 expression levels increase rapidly in the ischemic area after coronary artery ligation and remain elevated for several days in the perfused myocardium (Kempf loc. Cit.).

In a study with initially healthy women (Brown, D. A. et al., 2002, Concentration in plasma of macrophage inhibitory cytokine-1 and risk of cardiovascular events in women: a nested case-control study

[Concentration of plasma type 1 macrophage inhibitor cytokine and risk of cardiovascular events in women: a case study and nested controls], The Lancet, 359: 2159-2163) has shown that higher levels of GDF-15 predict episodes cardiovascular in this population. It has been revealed that the threshold level indicative of a high risk was 856 pg / ml of GDF-15. Similarly, US Patent 2003/0232385 reveals that GDF-15 levels greater than 850 pg / ml indicate an increased risk of cardiovascular events. Kempf et al. (Circulation, vol. 114, Suppl. S, p. 721), published in October 2006, demonstrate that circulating levels of GDF-15 are related to the severity of chronic heart failure and predict the risk of death.

Conventional techniques, especially in the case of urgent situations, do not normally serve to obtain a reliable diagnosis and / or to assess the risk. Therefore, based on these diagnostic techniques, a personalized treatment regimen cannot be established with sufficient accuracy. As a consequence, many patients will receive a treatment regimen that is insufficient or that may have adverse side effects. In many cases, acute cardiovascular events, once determined by the conventional diagnostic techniques mentioned above and / or by the patient's troponin levels, are usually treated by cardiac interventions. These cardiac interventions include several types of angioplastic operations and / or coronary bypass surgery, which are performed to restore the correct blood flow, eg inside the coronary vessels. However, these interventions are not always successful and may even be harmful to the patient. They are also long and expensive. The same difficulties and deficiencies of current risk assessment techniques are found in interventions for heart failure patients, eg in pharmacological therapies such as treatment with an angiotensin converting enzyme inhibitor, with a receptor blocker. angiotensin, with a beta-blocker or with an aldosterone antagonist, and in interventional therapies such as cardiac resynchronization therapy (CRT) or implantable cardioverter defibrillator (ICD).

Therefore, diagnostic or prognostic measures are needed that allow the stratification of the individual risk of a patient supposedly in need of a certain treatment regimen, such as a cardiac intervention.

The technical problem underlying the present invention can be interpreted as the provision of means and methods to meet the above needs.

Therefore, the present invention relates to a method for identifying a subject susceptible to cardiac intervention, which comprises

a) determine the amount of GDF-15 in a sample of a subject in need of cardiac intervention and

b) compare the amount of GDF-15 determined in step a) with a reference amount, in order to identify a subject susceptible to cardiac intervention.

The method of the present invention is preferably an in vitro method. It can also include additional steps to those explicitly indicated above. For example, the additional steps may consist of previous treatments of the sample or the evaluation of the results obtained by the method. The method of the present invention can also be used to control, confirm and subclassify a subject in need of cardiac intervention. The method can be carried out manually or with automatic assistance. Preferably step (a) and / or (b) can be fully or partially automated, eg by means of a suitably robotic sensor apparatus for the determination of stage (a) or a computer comparison in step (b).

As used herein, the term "identify" means assessing whether a subject is susceptible to cardiac intervention or not. Normally, as specialists in the field will understand, such assessment is not intended to be correct for all (that is, 100%) of the subjects identified. However, the term requires the identification of a statistically significant portion of subjects (eg a cohort in a cohort study). The person skilled in the art can determine whether or not a portion is statistically significant using several well-known tools of statistical evaluation, eg determination of confidence intervals, determination of p-value, t-Student test, Mann-Whitney test, etc. There are detailed descriptions in Dowdy and Wearden, Statistics for Research, John Wiley & Sons, New York 1983. The preferred confidence intervals are at least 90%, at least 95%, at least 97%, at least 98% or at least 99%. The p values are preferably 0.1, 0.05, 0.01, 0.005 or 0.0001. More preferably at least 60%, at least 70%, at least 80% or at least 90% of the subjects of a population can be correctly identified by the method of the present invention.

As used herein, the term "subject" refers to animals, preferably mammals and more preferably humans.

In any case, the above-mentioned method of the present invention contemplates that the subject is "in need of cardiac intervention", that is to say that it shows symptoms and / or physical signs clearly associated with an acute cardiovascular episode, eg chest pain, dyspnea, ECG and other changes, as described above. More preferably, the subject has one or more episodes of angina lasting at least 5 minutes in the previous 24 hours and a positive test for cardiac troponin T or I or at least 0-5 mm of transient or persistent depression ST segment not previously registered and not attributable to coexisting disorders. Alternative, but also preferably, that the subject shows symptoms of ischemia increasing or appearing at rest or justifying the suspicion of an acute myocardial infarction, with a last episode in the previous 48 hours. Cardiac ischemia should be verified by electrocardiography (ST depression = 0-1 mV or T wave inversion = 0-1 mV) or by increasing biochemical markers (creatine kinase [CK] -MB> 6 ug / l, troponin-T > 0.01 ng / ml, qualitative test of positive troponin-T or catalytic activity of CK, CK-B, or CK MB higher than the local diagnostic limit for myocardial infarction).

Acute cardiovascular events are preferably acute coronary syndromes (ACS). ACS patients may present with unstable angina pectoris (API) or myocardial infarction (MI). The IM can be with ST segment elevation (STEMI) or without elevation (NSTEMI). The onset of an ACS can be followed by left ventricular dysfunction (DVI) and symptoms of heart failure.

The term "cardiac intervention" encompasses treatment regimens that include surgery, microsurgery or other invasive therapies that affect the cardiovascular system and preferably the heart. As used herein, cardiac interventions are preferably treatment regimens whose objective is to restore the adequate supply of oxygen to the heart, which is preferably achieved by restoring blood flow through the blood vessels that support the heart, i.e. the vessels. coronary blood These blood vessels can be damaged due to thrombotic or atherosclerotic plaques. Accordingly, cardiac interventions will preferably comprise the destruction and / or removal of said plaques and, if necessary, vessel repair. Cardiac interventions according to the present invention are chosen from the group consisting of percutaneous coronary angioplasty, percutaneous transluminal coronary balloon angioplasty, laser angioplasty, coronary stent implantation or bypass implantation.

The term "sample" refers to blood, plasma or serum samples.

The term "growth differentiation factor-15" or "GDF-15" refers to a member polypeptide of the cytokine superfamily that includes transforming growth factors-β. The terms polypeptide, peptide and protein are used interchangeably throughout this description. GDF-15 was originally cloned as a type 1 macrophage inhibitor cytokine and was later also identified as a transforming growth factor-placental β, placental bone morphogenetic protein, a non-steroidal anti-inflammatory drug-1 gene and a prostate-derived factor (Bootcov loc cit; Hromas, 1997 Biochim Biophys Acta 1354: 40-44; Lawton 1997, Gene 203: 17-26; Yokoyama-Kobayashi 1997, J Biochem (Tokyo), 122: 622-626; Paralkar 1998, J Biol Chem 273: 1376013767). Similarly to other TGF-β-like cytokines, GDF-15 is synthesized as an inactive precursor protein that undergoes disulfide bridge homodimerization. After proteolytic dissociation of the N-terminal propeptide GDF-15 is secreted as a 28 kDa dimer protein (Bauskin 2000, Embo J 19: 2212-2220). The amino acid sequences of GDF-15 are disclosed in WO 99/06445, WO 00/70051, WO 2005/113585 Bottner 1999, Gene 237: 105-111, Bootcov loc. cit, Tan loc. cit., bask 2001, Mol Pharmacol 59: 901908, Hromas loc cit, Paralkar loc cit, Morrish 1996, Placenta 17: 431-441 or Yokoyama-Kobayashi loc cit .. The GDF-15 used here also comprises variants of the GDF polypeptides -15 specific mentioned above. These variants have at least the same essential biological and immunological properties as specific GDF-15 polypeptides. In particular, they share the same essential biological and immunological properties if they are detectable by the same specific analyzes cited in this description, eg by ELISA assays with polyclonal or monoclonal antibodies that specifically recognize said GDF-15 polypeptides. A preferred test is described in the attached examples. On the other hand it should be understood that a variant according to the present invention will have an amino acid sequence that differs by substitution, deletion and / or addition of at least one amino acid and that preferably is still identical in at least 50%, 60%, 70 %, 80%, 85%, 90%, 92%, 95%, 97%, 98% or 99% to the amino acid sequence of the specific GDF-15 polypeptides. The degree of identity between two amino acid sequences can be determined by well-known algorithms of the technical state. The degree of identity should preferably be determined by comparing two optimally aligned sequences through a comparative window that includes a fragment of the amino acid sequence with additions or deletions (eg gaps or bumps) with respect to the reference sequence (which does not include additions or deletions). The percentage is calculated by determining the number of positions of both sequences in which the same amino acid residue appears, dividing this number of matching positions by the total number of positions in the comparative window and multiplying the result by 100, to obtain the percent of sequential identity. Optimal sequence alignment for comparison can be performed using the local homology algorithm of Smith and Waterman Add. APL Math 2: 482 (1981), the homology alignment algorithm of Needleman and Wunsch J. MoL BioL 48: 443 (1970), the similarity search method of Pearson and Lipman Proc. NatL Acad Sci. (USA) 85: 2444 (1988), the computerized implementations of these algorithms (GAP, BESTFIT, BLAST, PASTA and TFASTA in the program package of Wisconsin Genetics, Genetics Computer Group (GCG), 575 Science Dr. , Madison, WI) or by visual inspection. If there are two sequences identified for comparison, GAP and BESTFIT are preferably used to determine their optimal alignment and therefore their degree of identity. Preferably, default values of 5.00 are used for the weight of the hole and 0.30 for the length of the hole. The variants mentioned above can be allelic or any other species of homologs, paralogs or specific orthologs. In addition, the variants mentioned herein include fragments of the specific GDF-15 polypeptides or the types of variants mentioned above, provided that these fragments have the essential immunological and biological properties mentioned above. Such fragments can be eg degradation products of GDF-15 polypeptides. Variants that differ by post-translational modifications such as phosphorylation or myristylation are also included.

Determining the amount of GDF-15 or any other peptide or polypeptide cited in this description means measuring the amount or concentration, preferably semiquantitatively or quantitatively. The measurement can be performed directly or indirectly. Direct measurement means measuring the amount or concentration of the peptide or polypeptide based on a signal obtained from the peptide or polypeptide itself, whose intensity is directly related to the number of peptide molecules present in the sample. Said signal - sometimes referred to herein as an intensity signal - can be obtained, for example, by measuring an intensity value of a specific physical or chemical property of the peptide or polypeptide. Indirect measurement consists in measuring a signal obtained from a secondary component (that is, from a component that is not the peptide or polypeptide itself) or from a biological data reading system, eg cellular responses, ligands, markers or Measurable enzyme reaction products.

According to the present invention the amount of a peptide or polypeptide can be determined by all known means to find out the amount of a peptide in a sample, including immunoassay devices and methods that use labeled molecules in various sandwich, competition or test formats. others. Such tests will produce a signal indicative of the presence or absence of the peptide or polypeptide. In addition, the signal strength can preferably be directly or indirectly related (eg inversely proportional) to the amount of polypeptide present in a sample. Other suitable methods consist of measuring a physical property

or specific chemistry of the peptide or polypeptide, such as its exact molecular mass or its NMR spectrum. Such methods preferably comprise biosensors, optical devices coupled to immunoassays, biochips, analytical devices such as mass spectrometers, NMR analyzers or chromatographic devices. Other methods are based on ELISA microplates, fully automated or robotic immunoassays (available for example in Elecsys® analyzers), CBA (an enzymatic cobalt fixation assay available for example in Roche-Hitachi® analyzers) and latex agglutination assays (available for example in Roche-Hitachi® analyzers).

The determination of the amount of a peptide or polypeptide preferably comprises the steps of (a) contacting a cell capable of eliciting a cellular response, the intensity of which indicates the amount of peptide or polypeptide, with said peptide or polypeptide for a period of time adequate, (b) measure the cellular response. To measure cellular responses, the sample, processed or not, is preferably added to a cell culture and an internal or external cellular response is measured. The cellular response may include the measurable expression of a reporter gene or the secretion of a substance, eg a peptide, a polypeptide or a small molecule. The expression or substance generates an intensity signal related to the amount of peptide or polypeptide.

The determination of the amount of a peptide or polypeptide preferably also comprises the step of measuring an intensity signal from the peptide or polypeptide contained in the sample. As described above this signal can be the intensity observed at a specific m / z variable of the peptide or polypeptide analyzed in a mass spectrum or a specific NMR spectrum of the peptide or polypeptide.

The determination of the amount of a peptide or polypeptide may preferably comprise the steps of (a) contacting the peptide with a specific ligand, (b) eliminating (optionally) the unbound ligand, (c) measuring the amount of fixed ligand . The fixed ligand generates an intensity signal. According to the present invention, the fixation can be covalent or non-covalent. According to the present invention a ligand can be any compound, eg a peptide, a polypeptide, a nucleic acid or a small molecule, which binds to the peptide or polypeptide described herein. Preferred ligands include antibodies, nucleic acids, peptides or polypeptides as receptors or fixers of the peptide or polypeptide and fragments thereof comprising the binding domains for the peptides, and aptamers, eg nucleic or peptide acids. The methods for preparing said ligands are well known in the technical state. For example, the identification and production of suitable antibodies or aptamers is also provided by commercial suppliers. The specialist in the field knows methods to create derivatives of these ligands that are more related or specific. For example, random mutations can be introduced into nucleic acids, peptides or polypeptides. The fixation of these derivatives can then be tested according to known selection procedures of the technical state, eg phage expression. The antibodies mentioned herein can be polyclonal and monoclonal, and also fragments thereof, such as Fv, Fab and F (ab) 2, capable of binding antigens or haptens. The present invention also includes single chain antibodies and humanized hybrid antibodies, in which amino acid sequences of a non-human donor antibody that possess a desired specificity against antigens and sequences of a human acceptor antibody are combined. Donor sequences will normally include at least the antigen-fixing amino acid residues, but may also comprise other structurally and / or functionally relevant amino acid residues of the donor antibody. Such hybrids can be prepared by several well known methods of the technical state. The ligand or agent preferably binds specifically to the peptide or polypeptide. Specific binding, according to the present invention, means that the ligand or agent should not substantially bind (by "cross reaction") to another peptide, polypeptide or substance existing in the analyzed sample. The specifically fixed peptide or polypeptide should preferably be linked with an affinity at least 3 times greater, preferably at least 10 times greater and especially at least 50 times greater than any other relevant peptide or polypeptide. A non-specific binding is tolerable if it can still be distinguished and measured unequivocally, for example, according to its size in a Western Blot,

or for its greater abundance in the sample. Ligand binding can be measured by any known method of the technical state. This method is preferably semiquantitative or quantitative. Suitable methods are described below.

Firstly, ligand fixation can be measured directly, eg by NMR or by surface plasmon resonance.

Secondly, if the ligand also serves as a substrate of an enzymatic activity of the peptide or polypeptide in question, a product of the enzymatic reaction can be measured (eg, the amount of a protease can be measured by determining the amount of substrate cleaved , eg in a Western Blot). Alternatively, the ligand itself may have enzymatic properties and the "ligand / peptide or polypeptide" complex or the ligand bound by the peptide or polypeptide, respectively, can be contacted with a suitable substrate that allows detection through the intensity signal generated. To measure enzymatic reaction products the amount of substrate should preferably be saturating. The substrate can also be labeled before the reaction with a detectable label. Preferably the sample is contacted with the substrate for a suitable period of time, understanding as such the time necessary to obtain a detectable amount of product, preferably measurable. Instead of measuring the quantity of product, the time required for a certain quantity of product to appear (eg detectable) can be measured.

Thirdly, the ligand can be covalently or non-covalently coupled to a marker that allows the ligand to be detected and measured. The ligand can be labeled by direct or indirect methods. Direct labeling involves coupling the marker directly (by covalent or non-covalent bond) to the ligand. Indirect labeling consists of joining (by covalent or non-covalent bonding) a secondary ligand to the first ligand. The secondary ligand should specifically bind to the first ligand. This secondary ligand can be coupled to an appropriate marker and / or be the target (receptor) of a tertiary ligand that binds to the secondary ligand. Secondary, tertiary or even higher order ligands are often used to increase the signal. Suitable secondary and higher order ligands may be antibodies, secondary antibodies and the well-known streptavidin biotin system (Vector Laboratories, Inc.). The ligand or substrate can also be "tagged" with one or more known tags of the technical state, which can be targets for other higher order ligands. Suitable labels include biotin, digoxigenin, His-Tag, glutathione-S-transferase, FLAG, GFP, myctag, hemagglutinin of influenza A (HA) virus, maltose binding protein and the like. In the case of a peptide or polypeptide the labeling is preferably at the N-terminal and / or C-terminal end. Suitable markers are all that can be detected by an appropriate method. Typical markers include gold particles, latex beads, acridane esters, luminol, ruthenium, enzymatically active markers, radioactive markers, magnetic markers ("eg magnetic beads", including paramagnetic and superparamagnetic markers) and fluorescent markers. Enzymatically active markers include eg horseradish peroxidase, alkaline phosphatase, beta-galactosidase, luciferase and derivatives thereof. Suitable substrates for detection include diaminobenzidine (DAB), 3,3'-5,5'-tetramethylbenzidine, NBT-BCIP (4-nitro-tetrazolium blue chloride and 5-bromo-4-chloro-3-indolyl- phosphate, available as a freshly prepared standard solution from Roche Diagnostics, CDP-Star® (Amersham Biosciences), ECF® (Amersham Biosciences) A suitable combination of enzyme-substrate can yield a color, fluorescent or chemiluminescent reaction product measurable by known methods of the technical state (eg, using a light-sensitive film or an appropriate camera system.) To measure the enzymatic reaction, the criteria indicated above are similarly used. Typical fluorescence markers include fluorescent proteins (such as GFP and its derivatives), Cy3, Cy5, Texas red, fluorescein and Alexa dyes (eg Alexa 568) Other fluorescent markers can be purchased eg from Molecular Probes (Oregon). quantum os as fluorescent markers. Typical radioactive markers include S35, I125, P32, P33 and the like. A radioactive marker can be detected by any known and appropriate method, eg by means of a photosensitive film or a phosphor photosensor "Phosphor Imager". Suitable measurement methods according to the present invention also include precipitation (in particular immunoprecipitation), electrochemiluminescence (electrically generated chemiluminescence), RIA (radioimmunoassay), ELISA (enzyme-linked immunosorbent assay), sandwich immunoenzymatic assays, electrochemiluminescent type immunoassays sandwich (ECLIA), lanthanide augmented dissociation fluoroimmunoassay (DELFIA), scintillation proximity test (SPA), turbidimetry, nephelometry, turbidimetry or nephelometry enhanced with latex or solid phase immunoassays. Other known methods of the technical state (such as gel electrophoresis, 2D gel electrophoresis, SDS polyacrylamide gel electrophoresis (SDS-PAGE), Western Blotting and mass spectrometry) can be used individually or in combination with markers or other methods of detection, as described above.

The amount of a peptide or polypeptide can also be determined as follows: (a) contacting a solid support carrying a ligand for the peptide or polypeptide, as described above, with a sample containing the peptide or polypeptide and (b) measure the amount of peptide or polypeptide attached to the support. The ligand, preferably chosen from the group consisting of nucleic acids, peptides, polypeptides, antibodies and aptamers, is preferably immobilized on a solid support. Materials for the production of solid supports are well known in the technical state and include, among others, commercially available column materials, polystyrene beads, latex beads, magnetic beads, colloidal metal particles, chips and glass and / or silicon surfaces. , nitrocellulose strips, membranes, sheets, dura, wells and walls of reaction plates, plastic tubes, etc. The ligand or agent can be attached to many different supports. As well-known examples of supports, glass, polystyrene, polyvinyl chloride, polypropylene, polyethylene, polycarbonate, dextran, nylon, amylases, natural and modified celluloses, polyacrylamides, agarose and magnetite can be mentioned. For the purposes of the present invention the support can be of the soluble or insoluble type. Suitable methods for fixing / immobilizing said ligand are well known and include, but are not limited to, ionic, hydrophobic, covalent and similar interactions. The use of "suspended matrices" as matrices according to the present invention is also contemplated (Nolan 2002, Trends BiotechnoL 20 (1): 9-12). In these suspended matrices, the support, eg a microbead or microsphere, is suspended. The matrix consists of different microbeads or microspheres, possibly marked, that carry different ligands. Methods for producing such matrices, based for example on solid phase chemistry and photolabile protecting groups, are generally known (US 5,744,305).

As used herein, the term "amount" encompasses the absolute amount of a polypeptide or peptide, the relative amount or concentration of said polypeptide or peptide, as well as any value or parameter related thereto or derivative thereof. Said values or parameters include the values of the intensity signals of all the specific physical or chemical properties obtained by direct measurements of said peptides, eg intensity values in the mass or NMR spectra. Also included are all the values or parameters obtained by indirect measurements specified in any part of this description, eg determined expression levels of the biological reading systems in response to the peptides or intensity signals obtained from the specifically bound ligands. It should be understood that the values related to the quantities or parameters mentioned above can also be obtained through all current mathematical operations.

As used herein, the term "compare" refers to the comparison of the amount of the peptide or polypeptide contained in the sample under analysis with an amount of a suitable reference source specified anywhere in this description. It should be understood that, as used herein, "comparing" refers to the comparison of corresponding parameters or values, eg, an absolute amount is compared with another absolute reference amount, while one concentration is compared with another reference concentration, or an intensity signal obtained from a tested sample is compared with the same type of intensity signal of a reference sample. The comparison mentioned in step b) of the method of the present invention can be performed manually or by computer. In the case of computer-assisted comparison, the quantitative value found can be compared by means of a computer program with values corresponding to suitable references stored in a data bank. In addition, the computer program can evaluate the result of the comparison, that is, automatically provide the desired assessment in an appropriate output format. Based on the comparison of the amount determined in step a) with the reference amount, it is possible to see if a subject can undergo cardiac intervention, that is, if it belongs to the group of subjects that can be treated successfully by cardiac intervention. Therefore, the reference amount should be chosen so that a difference or similarity between the compared amounts allows the subject analyzed to be identified as belonging to the group of subjects susceptible to cardiac intervention or to identify those analyzed subjects that cannot undergo cardiac intervention.

Thus, as used herein, the term "reference amount" refers to an amount that allows to assess, as indicated above, if a subject in need of cardiac intervention can undergo it. Therefore the reference may come from (i) a known subject that has been treated satisfactorily, that is, without the occurrence of adverse effects such as reinfarction or mortality or from side effects caused by the treatment regimen, or from (ii) a subject known that he has not been treated satisfactorily, that is, that he has suffered reinfarction or that he died from cardiovascular complications after cardiac intervention or that he does not obtain any benefit from the treatment regimen. In addition, the reference amount defines a limit value, above which a subject will be considered susceptible to cardiac intervention and below which it will be judged that a subject cannot be treated satisfactorily by cardiac intervention. The amount of reference applicable to each subject will vary depending on several physiological parameters, such as age, sex or subpopulation, as well as the means used for the determination of said polypeptide or peptide. By means of the method of the present invention, a suitable reference amount can be determined, from a reference sample to be analyzed together with the test sample, that is to say simultaneously or subsequently. As the reference quantity that serves as a threshold value, the upper normal limit (LNS) can be taken, that is, the upper limit of the physiological amount found in a population of apparently healthy subjects. The LNS of a given population of subjects can be calculated by various well known techniques. An appropriate technique may be to determine the population median for the amounts of peptide or polypeptide analyzed by the method of the present invention. A threshold value (that is, a reference amount) of the GDF-15 is at least once to twice the LNS. In this context the LNS is 1200 pg / ml.

Therefore the reference amount defining a threshold value of GDF-15 according to the present invention is 1200 pg / ml.

An amount of GDF-15 greater than the reference indicates that a subject is susceptible to cardiac intervention.

In the study underlying the present invention it has been found advantageously that GDF-15 is a reliable prognostic biomarker for assessing the success of cardiac interventions in subjects in need thereof, that is, subjects suffering from cardiovascular complications and in particular those who are affected by acute cardiovascular events or heart failure. Thanks to the present invention, risk / success stratification can be easily performed before subjecting a patient to cardiac intervention. If the patient is not suitable for cardiac intervention, this dangerous therapy can be avoided, which is time-consuming and expensive. Therefore, in addition to preventing the subject from the adverse and serious side effects that accompany cardiac intervention, the method of the present invention will be beneficial to the healthcare system by saving resources. According to the method of the present invention described herein, both above and hereinafter, it should be understood that the amount of GDF-15 or the means for its determination can be used in the elaboration of a diagnostic composition intended to identify a subject susceptible to a cardiac intervention

In a preferred embodiment of the present invention said method further includes determining the amount of a cardiac troponin in said subject sample and comparing the amount of cardiac troponin with a reference amount.

The term "cardiac troponins" refers to all troponin isoforms expressed in heart cells and preferably in subendocardial cells. These isoforms are well identified in the technical state, as described eg in Anderson 1995, Circulation Research, vol. 76, no. 4: 681-686 and in Ferrieres 1998, Clinical Chemistry, 44: 487-493. Cardiac troponin preferably refers to troponin T and / or troponin I and more preferably to troponin T. It should be understood that in the method of the present invention the troponin isoforms can be determined jointly, that is simultaneously or sequentially, or individually, that is, without analyzing the other isoform in any way. The amino acid sequences of human troponin T and troponin I are disclosed in Anderson, loc cit and Ferrieres 1998, Clinical Chemistry, 44: 487-493.

Therefore, the term "cardiac troponin" also includes variants of the specific troponins mentioned above, that is, preferably troponin T or troponin I. Such variants have at least the same biological and immunological properties as the specific cardiac troponins. Specifically, they share the same essential biological and immunological properties, if they are detectable by the same specific assays mentioned in this description, eg by ELISA assays with polyclonal or monoclonal antibodies that specifically recognize said cardiac troponins. It should also be understood that a variant, within the meaning of the present invention, will have an amino acid sequence that differs by substitution, deletion and / or addition of at least one amino acid and that preferably remains at least 50% identical, 60% , 70%, 80%, 85%, 90%, 92%, 95%, 97%, 98% or 99% to the amino acid sequence of the specific troponin. The variants can be allelic or any other species of homologs, paralogs or specific orthologs. In addition, the variants mentioned herein include fragments of the specific cardiac troponins or of the types of variants mentioned above, provided that these fragments have the essential immunological and biological properties mentioned above. Such fragments can be eg degradation products of troponins. Variants that differ by post-translational modifications such as phosphorylation or myristylation are also included.

As already stated above, a preferred reference amount that serves as a threshold can be taken from the LNS. The LNS for a given population of subjects can be set as specified elsewhere in this description. A preferred threshold (ie, a reference amount) of a cardiac troponin, in particular troponin T or I, is at least once, more preferably two to five times, the LNS. Preferably, in this context, the LNS is 0.01 ng / ml for troponin T and 0.1 ng / ml for troponin I.

Thus, the reference amount defining a threshold value for troponin T is preferably, according to the present invention, 0.01 ng / ml, 0.02 ng / ml or 0.05 ng / ml.

More preferably, an amount of cardiac troponin greater than the reference is indicative that a subject is susceptible to cardiac intervention.

In another preferred embodiment of the present invention the method also comprises (i.e., in addition to the determination of GDF-15 and / or cardiac troponin) the determination of the amount of a natriuretic peptide in said subject sample and the comparison of the amount of the natriuretic peptide with a reference.

The term "natriuretic peptide" includes the types called atrial natriuretic peptide (ANP) and cerebral natriuretic peptide (BNP) and variants thereof with the same predictive potential. Natriuretic peptides according to the present invention comprise ANP and BNP types and variants thereof (see eg Bonow, 1996, Circulation 93: 1946-1950). ANP peptides include the pre-proANP, proANP, NT-proANP and ANP types. BNP peptides include the pre-proBNP, proBNP, NT-proBNP and BNP types. The pre-pro-peptide (134 amino acids in the case of the pre-proBNP) comprises a short signaling peptide that enzymatically dissociates releasing the propeptide (108 amino acids in the case of proBNP). In turn, the propeptide dissociates into an N-terminal propeptide (NT-propeptide, 76 amino acids in the case of NT-proBNP) and the active hormone (32 amino acids in the case of BNP, 28 amino acids in the case of ANP).

Preferred natriuretic peptides according to the present invention are NT-proANP, ANP, NT-proBNP, BNP and variants thereof. ANP and BNP are the active hormones and have a shorter half-life than their respective inactive counterparts NT-proANP and NT-proBNP. BNP is metabolized in the blood, while NT-proBNP circulates in it as an intact molecule and as such is eliminated renally. The in vivo half-life of the NT-proBNP is 120 minutes longer than that of the BNP, which is 20 minutes (Smith 2000, J EndocrinoL 167: 239-46.). The preanalytics with NT-proBNP are more resistant and facilitate the transport of the sample to a central laboratory (Mueller 2004, Clin Chem Lab Med 42: 942-4.). Blood samples can be stored at room temperature for several days or can be sent by mail or issued without loss of recovery. In contrast, the storage of BNP for 48 hours at room temperature or at 4 ° C produces a concentration loss of at least 20% (Mueller loc.cit; Wu 2004, Clin Chem 50: 867-73.). Therefore, depending on the period of time or the properties of interest, it may be more convenient to measure the active forms or the inactive forms of the

natriuretic peptide

The most preferred natriuretic peptides according to the present invention are NT-proBNP or variants thereof. As described briefly above, human NT-proBNP, according to the present invention, is a polypeptide preferably comprising a length of 76 amino acids corresponding to the N-terminal portion of the human NT-proBNP molecule. The structure of the human BNP and NT-proBNP has already been described in detail in the prior technical state, eg in WO 02/089657, WO 02/083913 or in Bonow loc. cit. The NT-proBNP used herein is preferably human NT-proBNP as described in EP 0 648 228 B1. These documents of the prior technical state are incorporated herein by reference as regards the specific sequences of the NT-proBNP and its variants disclosed therein. The NT-proBNP according to the present invention also comprises alleles and other variants of said specific sequence of the human NT-proBNP discussed above. Specifically, polypeptide variants are contemplated which at the amino acid level are at least 60% identical, more preferably at least 70%, at least 80%, at least 90%, at least 95%, at least 98% or at least 99% to human NT-proBNP. Basically similar, proteolytic degradation products still recognizable by diagnostic means or by ligands directed against the corresponding full-length peptide are also contemplated. Also included are variants of polypeptides with deletions, substitutions and / or additions of amino acids with respect to the amino acid sequence of the human NT-proBNP, provided that said polypeptides have NT-proBNP characteristics, which in this case refer to immunological properties and / or Biological Variants of NT-proBNP preferably have immunological properties (ie epitopic composition) comparable to those of NT-proBNP. Thus the variants will be recognizable by the aforementioned means or ligands used to determine the amount of natriuretic peptides. The biological and / or immunological properties of NT-proBNP can be detected by the assay described in Karl et al. (Karl As already stated above, a preferred reference amount that serves as a threshold can be taken from the LSN. The LSN for a population specific subjects can be determined in the manner specified in any part of this description A preferred threshold value (ie, a reference amount) for a natriuretic peptide and in particular for the NT-proBNP is at least once, with greater preference two to four times, the LSN In this context the LSN of the NT-proBNP is preferably 300 pg / ml The LSN of other natriuretic peptides known in the technical state are 40 pg / ml for the ANP, 100 pg / ml for the BNP and 500 pmol / l for the NT-proANP.

Therefore, the reference amount defining a threshold value of NT-proBNP according to the present invention is preferably 600 pg / ml or 1200 pg / ml and more preferably 1000 pg / ml.

More preferably, an amount of a natriuretic peptide greater than the reference is indicative that a subject is susceptible to cardiac intervention.

It should be understood that the definitions and explanations of the terms indicated above and hereafter apply correspondingly throughout the remaining statement.

A device is also disclosed for identifying a subject susceptible to cardiac intervention, adapted to the performance of the method of the present invention, comprising means for determining the amount of GDF-15 in a sample of the subject and means for comparing said amount with one of reference, identifying a subject susceptible to cardiac intervention.

As used herein, the term "device" refers to a system comprising at least the aforementioned means, operatively connected to each other to allow prediction. The preferred means for determining the amount of GDF-15, preferably in combination with a cardiac troponin and / or with a natriuretic peptide, and the means for making the comparison are disclosed above in relation to the method of the present invention. The way of connecting the media in an operative way will depend on the type of media included in the device. For example, where means are used to automatically determine the amount of peptides, the data collected by these means of automatic operation can be processed, eg by a computer program, to obtain the desired results. In such a case, the means are preferably comprised in a single device, which, therefore, can include an analyzer unit for measuring the amount of peptides or polypeptides in an applied sample and a computer unit for processing the data resulting from the evaluation. . Alternatively, where means such as test strips are used to determine the amount of peptides or polypeptides, the comparison means may comprise control strips or tables that allocate the determined amount to a reference amount. The test strips are preferably coupled to a ligand that specifically binds to said peptides or polypeptides. The strip or device preferably includes means for detecting the binding of said peptides or polypeptides to said ligand. Preferred detection means are disclosed in relation to the method of the present invention described above. In this case, the means are operatively connected so that the system user gathers the result of the determination of the quantity and its diagnostic or predictive value, following the instructions and interpretations indicated in a manual. In said embodiment, the means may be presented as separate devices and, preferably, packaged together in kit form. The specialist in the field will easily recognize how to connect the media. Devices that can be applied without the expertise of a clinical physician, such as test strips or electronic devices that should only be loaded with a sample, are preferred. The results may be indicated as raw data that should be interpreted by the doctor. However, preferably, the output of the device is processed data, ie evaluated, whose interpretation does not require a doctor. Other preferred devices are the analysis units / apparatus (eg biosensors, matrices, solid supports coupled to ligands that specifically recognize the natriuretic peptide, surface plasmon resonance devices, NMR spectrometers, mass spectrometers, etc.) or evaluation units / apparatus such as those mentioned above in accordance with the present invention.

An embodiment kit of the method of the present invention for identifying a subject susceptible to cardiac intervention is also disclosed. Said kit comprises means for determining the amount of GDF-15 in a sample of a subject and means for comparing said amounts with other references, identifying a subject susceptible to cardiac intervention.

As used herein, the term "kit" refers to a set of the aforementioned means, preferably available separately or in a single package, also preferably comprising instructions for carrying out the method of the present invention.

The figures represent:

Figure 1: (A) Cumulative probability of acute myocardial infarction for 2 years by tertiles of GDF-15 levels at the admission of 1034 non-invasive patients with ACS-NSTE (<1200 n = 400, 1200 - 1800 n = 394,> 1800 n = 240) enrolled in the FRISC II trial (P <0.0001 per logarithmic range test). (B) Cumulative probability of acute myocardial infarction for 2 years by tertiles of GDF-15 levels at the admission of 1045 invasive patients with ACS-NSTE (<1200 n = 416, 1200 - 1800 n = 376,> 1800 n = 253 ) enrolled in the FRISC II trial (P = 0.6915 per logarithmic range test).

Figure 2: Cumulative probability of death for 1 year according to tertiles of GDF-15 levels at the admission of 2081 patients with ACS-NSTE enrolled in the GUSTO-IV trial (P <0.001 per logarithmic range test).

Figure 3: Analysis by curves of the receiver's operational characteristic (ROC) that relate biomarker levels to mortality during 1 year in 2081 patients with ACS-NSTE enrolled in the GUSTO-IV trial. The areas calculated below the curves were 0.757 for, 0.735 for NT-proBNP, 0.728 for creatinine removal, 0.629 for CRP and 0.620 for troponin T.

Figure 4: Mortality during 1 year of follow-up between subgroups of patients with ACS-NSTE enrolled in the GUSTO-IV trial, without (panel A) and with (panel B) a previous history of myocardial infarction, according to tertiles of GDF levels -15 and NT-proBNP upon entry. Each column indicates the number of deaths by number of patients.

Figure 5 (not part of the present invention). (A) Cumulative survival in patients with chronic heart failure of the GDF-15 bypass quartile cohort (P <0.001 per logarithmic range test). (B) Cumulative survival in patients with chronic heart failure of the validation cohort; Patients were grouped according to the cut-off levels defined in the shunt cohort (P <0.001 per logarithmic range test). The number of patients in danger is indicated below.

Figure 6 (not part of the present invention). Univariate predictors of mortality during the follow-up of patients with chronic heart failure of the bypass (A) and validation (B) cohorts. Risk indices with 95% confidence intervals, Chi2 values and P values are indicated. Creatinine, uric acid, NT-proBNP and GDF-15 were not normally distributed and therefore transformed into their natural logarithms before of the analysis; Risk indices refer to an increase of one unit in the scale of natural logarithms of these variables.

Figure 7 (not part of the present invention). (A) Survival index without adverse episodes of patients over a period of 30 days. Patients have been classified according to their GDF-15 levels (higher or lower than 4,600 ng / l); (B) Survival index over a period of 180 days.

The following examples illustrate exclusively the present invention. In no way should they be construed as limiting the scope of the present invention.

Example 1: Determination of GDF-15, NT-proBNP and troponin in serum and plasma samples

To determine the concentration of GDF-15 in serum and plasma samples an immunoradiometric assay (IRMA) was developed, using a goat anti-human GDF-15 polyclonal IgG antibody, purified by affinity chromatography with GDF-15, from R&D Systems ( AF957). At night, Maxisorp Startubes tubes (Nunc) were coated at 4 ° C with 0.5 µg of anti-GDF-15 IgG in 0.1 M Na carbonate buffer (pH 9.0) and then washed twice with physiological serum containing 0.1% Tween 20, phosphate buffered. Serum or plasma samples (100 µl) were diluted 1: 1 with assay buffer (30 g / l BSA, 10 g / l bovine IgG, 1% goat serum, 0.1% Na azide, 1M NaCl, 40 mM Na phosphate buffer, pH 7.4), were added to the tubes and incubated for 16 hours at 4 ° C. After two washing steps, 10 ng of iodinated anti-GDF-15 IgG [I125] (specific activity 0.74 MBq / µg) was diluted in 200 µl of assay buffer, added to each tube and incubated for 4 hours at room temperature. After three final washing steps the fixed radioactivity was quantified in a gamma counter (LKB Wallac 1261). A standard curve with recombinant human GDF-15 from R&D Systems (957-GD / CF) was generated in each test. The results with new batches of recombinant GDF-15 protein were checked in standard plasma samples and any deviation greater than 10% was corrected by introducing an adjustment factor for this assay. GDF-15 measurements in serum and plasma samples from the same patient gave practically identical results after final correction for dilution factors. The detection limit of the assay was 20 pg / ml. The intra-assay coefficient of variation determined for average GDF-15 levels of 744, 1518 and 8618 pg / ml was respectively 5.6, 5.9 and 6.5%. The intra-assay coefficient of variation determined for average GDF-15 levels of 832, 4739 and 9230 pg / ml was respectively 8.6, 5.7 and 4.4%.

Troponin T levels were determined by a third generation assay on an Elecsys 2010 analyzer (Roche Diagnostics), with a detection limit of 0.01 ng / ml.

NT-proBNP levels were determined by an immunoassay in an Elecsys 2010, with a detection limit of 20 pg / ml.

Example 2: Analysis of the differences between invasive and non-invasive treatment regimens in patients with ACS-NSTE from the FRISC II study

Patients for the FRISC II study were recruited between June 1996 and May 1998 in 58 Scandinavian hospitals, of which 16 were intervention centers. Patients could be chosen if they had symptoms of ischemia that increased or appeared at rest or justified the suspicion of acute myocardial infarction, with the last episode within 48 hours before the start of treatment with dalteparin or with standard heparin. Myocardial ischemia had to be verified by electrocardiography (ST depression ≥ 0-1 mV or T wave inversion ≥ 0-1 mV) or by increasing biochemical markers (creatine kinase [CK] -MB> 6 ug / l, troponin- T> 0.01 ng / ml, qualitative test of positive troponin-T or catalytic activity of CK, CK-B, or CK MB higher than the local diagnostic limit for myocardial infarction). The exclusion criteria were: increased risk of bleeding episodes, anemia

or indication or treatment of thrombolysis in the previous 24 hours, angioplasty in the last 6 months, be on a waiting list for coronary revascularization, another acute or severe heart disease, renal or hepatic insufficiency, osteoporosis recognized as clinically relevant, other serious diseases , hypersensitivity to randomized drugs, assumption of difficulties to cooperate or participate in this or other clinical trials, patients with previous open heart surgery, old age (eg> 75 years) or other disorders that made randomization inappropriate for revascularization early The FRISC II study was a prospective, randomized and multicenter trial with parallel groups. We compare invasive and non-invasive treatments by factorial design. Half of the patients in each group were randomly assigned long-term treatment with dalteparin or placebo subcutaneously for 3 months. The comparison of invasive and non-invasive strategies was open and the comparison of long-term dalteparin treatment with placebo was double blind.

In the invasive groups the objective was to perform all the invasive procedures within 7 days of the beginning with dalteparin exposed. The direct invasive treatments were: coronary angiography a few days after recruitment, with the intention of revascularizing within 7 days after the start of open treatment. Revascularization was recommended for all patients with an obstruction of at least 70% of the diameter in any artery that supplies a substantial portion of the myocardium. Percutaneous coronary intervention was recommended if there were one or two objective lesions and coronary artery bypass surgery was preferred in patients with three-vessel disease or the left main artery.

Non-invasive treatment included coronary angiography in patients with refractory or recurrent symptoms despite maximum medical treatment or severe ischemia in a test of limited effort due to symptoms, before discharge. The stress test criteria for performing angiography and revascularization were: ST depression ≥ 0.3 mV, limiting chest pain associated with a low maximum workload (<90 W in men or <70 W in women) or a decrease of blood pressure, or an ST elevation without previous concomitant Q waves or an inversion of the T wave in the stress test. During the follow-up, invasive procedures, regardless of the randomized strategy, were taken into account for all patients with disabling symptoms, recurrent instability or reinfarction.

Upon admission, patients were initially treated openly with subcutaneous infusion of dalteparin or standard heparin, adjusted for activated partial thromboplastin time. From the random distribution, all patients received 120 IU / kg of dalteparin every 12 hours (maximum dose of 10,000 IU) subcutaneously, for at least 5 days in the non-invasive group and until the end of the procedures in the invasive group The patients then received two daily subcutaneous injections of dalteparin or placebo. Women weighing less than 80 kg and men weighing less than 70 kg received 5,000 IU of dalteparin or placebo and men and women weighing more than these values received 7,500 IU. This regimen was maintained for 3 months, with patients self-injecting with pre-filled single dose syringes, after hospital discharge. The last injection of dalteparin treatment in an open or double-blind trial was administered before 12 h prior to coronary procedures. After angioplasty the administration of dalteparin was resumed

or placebo 2-6 hours after removing the sheath. After administration of an infusion of the abciximab glycoprotein IIb / IIIa inhibitor, administration of dalteparin or placebo was not resumed until 24 h later. After coronary artery bypass surgery, all patients received 5,000 IU of dalteparin twice daily, until mobilization, and double-blind treatment was initiated a few days before discharge. Compliance with this regimen was monitored by asking patients to record all injections in diaries and counting syringes returned or not used by outpatients. Upon admission, an initial dose of 300-600 mg of aspirin was administered to all patients, followed by a maintenance dose of 75-320 mg once daily. A β-blocker was administered, unless contraindicated. If necessary, organic nitrates and calcium antagonists could be added. According to modern treatment guidelines, the reduction of cholesterol with statins, angiotensin converting enzyme inhibitors for left ventricular dysfunction and aggressive antidiabetic treatment was recommended. The use of abciximab was encouraged during percutaneous coronary interventions. Ticlopidine was recommended for 3-4 weeks after stenting. Upon admission, or at the latest at random distribution, blood samples were analyzed locally to determine hemoglobin concentrations, leukocyte and platelet count, prothrombin time, creatinine, glucose, hemoglobin A10 if necessary, triglycerides , cholesterol, HDL and LDL cholesterol. The biochemical markers of myocardial injury were analyzed at admission, after new episodes of severe pectoral pain and before and 424 h after revascularization. The most frequently used myocardial lesion marker was CK-MB mass, but some centers used the activity of total CK, CK-B or both. In most hospitals, quantitative determination of troponin-T was available. For exploratory purposes we provide all centers with a qualitative test of troponin-T, the second generation of Cardiao-T (Roche-Boehringer Mannheim, Mannheim, Germany). At random distribution, blood samples from all patients were stored frozen at -70 ° C for central analysis of troponin-T and other markers. Conventional 12-lead electrocardiography was performed upon admission, at random distribution, within 24 hours prior to invasive procedures, upon discharge from the hospital, at 3 and 6 months in outpatient visits and in case of suspicion of

5 recurrent unstable angina or myocardial infarction. Patients in the non-invasive group who did not have contraindications did a limited effort test for symptoms by bicycle 16 hours before discharge. Electrocardiography was performed with a standard evaluation of left ventricular function in 1951 patients before discharge and always before invasive procedures. All results of the stress tests and electrocardiograms were sent to a central laboratory for evaluation.

10 The main objective was to verify the effects of the invasive and non-invasive strategy on the assessment criteria composed of death and myocardial infarction at 6 months. Other predefined criteria were total death, myocardial infarction, angina symptoms, need for subsequent coronary angiography and revascularization, bleeding and stroke episodes. Myocardial infarction was defined by two of the three typical conventional criteria: chest pain, electrocardiographic diagnostic recording (especially of new Q waves) or an increase in

15 biochemical marker of myocardial injury according to the following definitions. For myocardial infarction not related to the procedure: concentration of CK-MB mass greater than the diagnostic limit of the local hospital in one measurement; CK, CK-B or CKMB catalytic activity greater than double the local limit in a measurement. For myocardial infarction related to percutaneous coronary interventions: CK-MB mass 1.5 times the local hospital's diagnostic limit for myocardial infarction in one measurement; CK, CK-B or CKMB catalytic activity in

20 one measurement three times the limit or in two measurements 1.5 times the local limit. The new Q waves were only used for the diagnosis of myocardial infarction related to coronary artery bypass surgery. The causes of death should be established by necropsy. All reports of deaths, myocardial infarctions, increases in biochemical markers related to percutaneous coronary interventions and new Q waves in electrocardiography of the central laboratory were judged by an independent clinical acts committee. The

The quality of the data was ensured through the continuous verification of data collection in all case reports registered by external controllers employed by the sponsoring pharmaceutical company. All cardiac episodes (efficacy criteria) and adverse event data were continuously sent from the centers to the data and safety control board. The study complied with the Declaration of Helsinki and all local ethical committees approved the protocol.

30 The concentrations of GDF-15 in serum or plasma were determined upon admission, using the same immunoradiometric assay (IRMA) described in the GUSTO IV study (see previous example 1). The different proportions of transcendent episodes (myocardial infarction at 2 years, death at 2 years, myocardial infarction and death at 2 years) were recorded by tertiles of GDF-15 levels (<1200 pg / ml; 1200- 1800 pg / ml;> 1800 pg / ml) in the group assigned to the invasive treatment strategy and in the group assigned to the strategy of

35 non-invasive treatment. The Kaplan-Meier method was used to illustrate the temporal succession of episodes during follow-up with respect to the tertiles of GDF-15 in both groups (Figure 1).

The following table 1 shows the risks of mortality (death) and / or other acute myocardial infarction (AMI).

Table 1: 2-year result in relation to an invasive strategy versus a non-invasive strategy in 1045 invasive patients and 1034 non-invasive patients with ACS-NSTE enrolled in the FRISC II trial 40

  Non-invasive

invasive

 GDF-15 Tertiles

% of episodes (total N) % of episodes (total N) Estimated odds ratio (95% CI) P value

Death / AMI

<1200 1200-1800> 1800 9.3 (400) 16.5 (394) 27.9 (240) 9.6 (416) 11.2 (376) 14.6 (253) 1,044 (0.652 to 1,669) 0.637 (0.420 to 0.966) 0.442 (0.282 to 0.693) 0.8584 0.0337 0.0004

YO SOY

<1200 1200-1800> 1800 8.0 (400) 12.9 (394) 21.7 (240) 8.2 (416) 9.8 (376) 9.5 (253) 1,024 (0.619 to 1,693) 0.734 (0.468 to 1,150) 0.399 (0.225 to 0.638) 0.9278 0.1772 0.0003

Death

<1200 1200-1800> 1800 1.8 (400) 4.1 (394) 13.3 (240) 1.4 (416) 2.1 (376) 6.3 (253) 0.822 (0.274 to 2.466) 0.514 (0.217 to 1.215) 0.439 (0.234 to 0.823) 0.7261 0.1292 0.0102

Example 3: Analysis of mortality and short-term and long-term myocardial reinfarction in patients with ACS-NSTE from the GUSTO-IV study

The samples investigated were from patients with ACS-NSTE, from the GUSTO-IV trial conducted between 1999 and 2000. The

45 detailed design and main test results have been published (Simoons 2001, Lancet 98: 351-360; Ottervanger 2003, Circulation, 107: 437-442). The patients eligible for the study were at least 21 years of age and had had one or more episodes of angina lasting at least 5 minutes, within 24 hours after admission, and a positive on the cardiac troponin test or at minus 0.5 mm of ST segment depression. Patients were randomly treated with infusion of abciximab or placebo for 24 or 48 hours, in addition to standard medical treatment. In 30 days of follow-up, total mortality and the proportion of myocardial infarctions were recorded. After 1 year of follow-up, only total mortality was recorded. In a substudy with 399 consecutive patients recruited in Swedish locations, plasma samples were available upon admission (i.e. at baseline) and at 24 hours, 48 hours and 72 hours, which were the first ones analyzed in the present study. Based on the significant results of this cohort, it was decided to analyze GDF-15 levels in serum samples taken from another 1682 randomized patients upon entering the trial, in order to get at least 2000 patients and to investigate the interactions with other biomarkers used to the prognosis, that is troponin T, NT-proBNP, CRP and creatinine elimination.

Plasma samples were also obtained from apparently healthy subjects included in the group of Swedish women and men and in the study of ischemic heart disease (SWISCH). This population was made up of 429 elderly people who were matched by age and sex to another contemporary SCA-NSTE population included in the Fragmin trial and rapid revascularization during instability in coronary artery disease (FRISC II). The detailed design and some biomarker results of the SWISCH trial are published. Subjects with an abnormal 12-lead ECG at rest, cardiovascular medication, proven cardiovascular disease, another chronic disease or an acute disease were excluded from the control population. All participants in the SWISCH study were required to have normal levels of creatinine, blood glucose and hemoglobin, and normal white blood cell and platelet counts.

The characteristics of the baseline state are presented in the form of numbers and proportions. Continuous data are expressed in the form of median and interquartile ranges. Comparisons of continuous variables between cases and controls were evaluated by non-parametric Mann-Whitney U trials. To assess the relationships between GDF-15 levels and baseline state characteristics, and levels of troponin T, NT-proBNP, CRP and creatinine clearance, in the tertiles of GDF-15, both correlation coefficients were used Spearman's rank between GDF-15 and these factors, such as the Cochran-Armitage trend test for the relationships between the proportions of various groups. Spearman's correlation coefficients and the Wilcoxon assigned range test were used to compare the variations of GDF-15 over time in the patient group. Differences in proportions in the transcendent episodes (myocardial infarction at 30 days, death at 30 days, myocardial infarction and death at 30 days, death at 1 year) in tertiles of GDF-15 levels were assessed by Cochran-Armitage trend test. The Kaplan-Meier method was used to illustrate the temporal succession of episodes during follow-up with respect to the tertiles of GDF-15 and the statistical evaluation was performed using the logarithmic range test (Figure 2). To identify the predictors of 1-year death, simple logistic regression analyzes were used. All variables were then analyzed in a multiple logistic regression analysis. For the additional comparison of the prognostic values of GDF-15, troponin T, NT-proBNP, CRP and creatinine elimination with respect to 1-year death, receiver operating characteristic curves (ROCs) were generated and the areas under the curves (figure 3). All data analyzes were performed using the statistical software SAS 9.0.

GDF-15 levels in apparently healthy control subjects

The controls consisted of 288 men (67.1%) and 141 women (32.9%) with a median age of 65 years (interquartile range of 59 to 71 years), of which 14.5% were smokers usual. The median level of GDF-15 in this population was 762 pg / ml, with 460 pg / ml and 1191 pg / ml as the 10th and 90th percentiles, respectively. The upper normal limit (LNS) was therefore rounded to 1200 pg / ml. The median (interquartile range) of the levels of NT-proBNP, CRP and creatinine clearance was 74 (46 to 113) pg / ml, 1.40 (0.79 to 2.40) ug / ml and 73 (62 up to 86) ml / min, respectively. GDF-15 levels correlated directly with age (Spearman's ro = 0.21; P <0.001) and anti-inflammatory activity (CRP, ro = 0.18; P <0.001) and inversely with creatinine clearance ( ro = -0.14; P = 0.002). In this group there were no significant correlations regarding sex, smoking or NT-proBNP.

GDF-15 levels in SCA-NSTE patients upon admission

SCA-NSTE patients consisted of 1315 men (63.2%) and 766 women (36.8%) with a median age of 66 years (interquartile range of 57 to 74 years). Although the patients were somewhat older than the healthy controls (P = 0.014), the two cohorts were very comparable in terms of age and sex. Random treatment with abciximab had no influence on the GDF-15 levels measured at any time and therefore the randomly distributed groups were combined. Table 2 shows the onset of symptoms until admission, the prevalence of cardiovascular risk factors, previous manifestations and treatments of cardiovascular disease, signs of persistent ischemia on the ECG and basal levels troponin T, NTproBNP , CRP and creatinine elimination in this patient population. SCA-NSTE patients presented significantly (P <0.001) higher levels of GDF-15 compared to healthy controls; the median was 1445 pg / ml and the 10th, 33rd, 66th and 90th percentiles were respectively 850 pg / ml, 1187 pg / ml, 1817 pg / ml and 3314 pg / ml. Thus, about two thirds of patients with ACS-NSTE had GDF-15 levels above the LNS (90th percentile) of healthy controls. As this LNS corresponded to the lower tertile of the patients, they were stratified into tertiles (with cut-off limits of 1200 and 1800 pg / ml) when related to the result.

The increase in GDF-15 tertiles upon admission was directly related to age, female sex, history of hypertension and diabetes, previous manifestations of heart disease, ie angina pectoris, myocardial infarction, coronary revascularization and failure. cardiac, ACE inhibitory therapy, and also with markers 5 of persistent ischemia and necrosis, myocardial dysfunction and inflammation, as indicated by ST segment depression and troponin T, NTproBNP and CRP levels (table 2); GDF-15 levels were inversely related to smoking and creatinine elimination (table 2). In a multiple regression analysis, using the natural logarithm of GDF-15 as the dependent variable, the following factors were significantly related to the GDF-15: age (P <0.001), male sex (P <0.001), time since appearance of symptoms until admission

10 (P = 0.006; inverse relationship), smoking (P <0.001), diabetes (P <0.001), history of heart failure (P <0.001), ST segment depression (P = 0.050), NT-proBNP ( P <0.001), CRP (P <0.001) and creatinine elimination (P <0.001; inverse relationship). There was no independent relationship with respect to troponin T level (P = 0.436).

Table 2: Characteristics of patients with ACS-NSTE according to the tertiles of GDF-15 levels upon admission.

 Referral study Validation study Value P (n = 235) (n = 220)

Clinical characteristics Age [years] 66 (56-72) 63 (58-69) 0.161 Male sex 215 (91.5) 197 (89.5) 0.479 BMI [kg / m2] at 25.4 (23.0 - 29 , 3) 26.1 (23.7 - 28.7) 0.444 Ischemic etiology 168 (71.5) 140 (63.6) 0.085 Severity of heart failure and biomarkers NYHA Class 2.7 ± 0.8 2.2 ± 0.7 <0.001 I 12 (5.1) 31 (14.1) II 84 (35.7) 111 (50.5) III 96 (40.9) 72 (32.7) IV 43 (18.3 ) 6 (2,7) LVEF [%] 30 (24-40) 33 (27-38) 0.082 Creatinine [µmol / l] 102 (84-137) 99 (90-115) 0.430 Creatinine removal [ml / min ] a 71.5 (45.6 - 90.8) 68.4 (53.8 - 88.8) 0.638 Uric acid [µmol / l] b 422 (345 - 529) 390 (330 - 450) 0.392 Hemoglobin [ g / dl] c 13.8 (12.3-14.9) 14.0 (13.1-14.8) 0.108 NT-proBNP [ng / l] 1340 (434-3740) 521 (209-1070) <0.001 GDF-15 [ng / l] 2240 (1232 - 4010) 1465 (1004 - 2194) <0.001 Medication at baseline <0.001 ACE inhibitor or BRA 183 (77.9) 213 (96.8) 0.004 Blocker β 112 (47.7) 146 (66.4) <0.001 Diuretic 167 (71.1) 194 (88.2) <0.001 Spironolacton at 58 (24.7) 50 (22.7) 0.703 Values are n (%), median (25th and 75th percentiles) or mean ± SD. aData of 216 patients in the referral cohort and 198 patients in the validation cohort; bdatos of 207 patients in the validation cohort; data of 214 patients in the validation cohort. BMI means body mass index; LVEF, left ventricular ejection fraction; BRA, angiotensin receptor blocker.

fifteen

Temporal evolution of GDF-15 levels in the SCA-NSTE

The temporal evolution of serum GDF-15 levels during an episode of unstable coronary artery disease was studied in a cohort of 399 patients of the SCA-NSTE population, whose samples were taken upon admission and at 24 hours, 48 hours and 72 hours. Linear regression analysis of the natural log of GDF-15 versus 20 at the time revealed a weak but statistically significant slope (P = 0.010), which indicates that during this time interval there was a gradual rise. However, as shown in Table 3, the levels of GDF-15 at these four points were within the same range. The temporal variation between the results was very limited, as evidenced by the close correlation between the levels of GDF-15 at the time of entry and later (Table 3). 67.4%, 69.7% and 70.4% of the patients had GDF-15 levels higher than the LNS at 24 hours,

25 48 hours and 72 hours, respectively.

Tracing

Baseline 24 hours 48 hours 72 hours GDF-15 (ng / l) Medium 1499 1575 1630 1664 Interquartile range (1151 to 2203) (1112 to 2286) (1163 to 2396) (1140 to 2357) Spearman correlation coefficients - -0.89 0.80 0.72 Follow-up against baseline P <0.001 P <0.001 P <0.001 Variation of GDF-15 level --- 48 74 124 (follow-up against baseline, ng / l) Medium ( -111 to 253) (-146 to 360) (-143 to 434) Interquartile range P = 0.001 P <0.001 P <0.001

In a cohort of 399 patients, GDF-15 levels were determined upon admission (baseline) and at 24 hours, 48

5 hours and 72 hours. Spearman's correlation coefficients and P values were calculated to describe the relationships between GDF-15 levels in the follow-up and baseline. Variations in the levels of GDF15 in the follow-up, compared with those in the baseline, were evaluated by the Wilcoxon assigned range test.

GDF-15 levels and mortality in the SCA-NSTE

10 The risk of death of patients with ACS-NSTE during follow-up increased markedly at higher levels of GDF-15 upon admission (Figure 2). The Kaplan-Meier mortality curves in the different tertiles showed a rapid separation, with 30-day mortality rates of 0.6%, 2.0% and 4.3%, respectively (P <0.001). The separation between the curves continued during the first year after the incident and after 1 year of follow-up the mortality rates were 1.5%, 5.0% and 14.1% in the respective tertiles (P <0.001). ROC analyzes too

15 demonstrated that GDF-15 is a powerful biochemical indicator of mortality, with an area under the curve (ABC) of 0.757 compared to NT-proBNP (ABC = 0.735), creatinine elimination (ABC = 0.728), CRP (ABC = 0.629) and troponin T (ABC = 0.620) (Figure 3). By simple logistic regression analysis, age, history of hypertension, diabetes, angina pectoris or previous myocardial infarction, history of congestive heart failure and levels of troponin T, NT-proBNP, CRP, elimination of creatinine and GDF-15 with mortality at 1 year (table 4).

20 Using a multiple logistic regression method, age, previous myocardial infarction and the highest levels of NT-proBNP and GDF-15 proved to be the only independent predictors (Table 4). Among these independent risk indicators the level of GDF-15 proved to be the best predictor of mortality (table 4). The results did not change when using a method of progressive elimination or stratification of GDF-15 and the other biomarkers in tertiles. Similarly, by adding only one at a time in the framework of a gradual selection, the only variables

25 significant were age, previous myocardial infarction, NTproBNP and GDF-15. Specifically, the GDF-15 levels measured in the cohort of 399 patients upon admission and at a later time provided similar information on the 1-year mortality prognosis, although, apparently, the GDF-15 levels in the first 24 hours had the highest predictive value (table 5).

Table 4: Logistic regression analysis of 1-year mortality in 2081 patients with ACS-NSTE with respect to the 30 characteristics of the baseline, medical history and measurements upon admission

Univariate model Multivariate model

Estimated rate of

P value  Estimated rate of P value

possibilities (95% CI)

possibilities (95% CI)

Age (per year)

1,087 (1,066 to 1,108) <0.001 1,044 (1,013, to 1,077) 0.006

Sex (male vs. female)

0.723 (0.513 to 1.018) 0.064 1,134 (0.730 to 1,764) 0.574

Time from symptoms (hours) Smoking habit1 History of hypertension1 History of hypercholesterolemia1 Diabetes mellitus1 Previous angina pectoris1 Previous myocardial infarction1 Prior revascularization1 History of heart failure1 ST segment depression  0.5 mm1 Troponin T2 NT-proBNP2 CRP2 Elimination creatinine3 GDF-152

0.995 (0.971 to 1.020) 0.727 (0.466 to 1,135) 1,796 (1,265 to 2,550) 1,016 (0.692 to 1,490) 2,251 (1,564 to 3,240) 1,727 (1,221 to 2,445) 2,945 (2,087 to 4,156) 0.746 (0.437 to 1,273) 3,404 ( 2,124 to 5,454) 1,348 (0,851 to 2,134) 1,255 (1,141 to 1,381) 1,902 (1,674 to 2,161) 1,453 (1,278 to 1,654) 0.962 (0.954 to 0.971) 4,817 (3,625 to 6,402) 0.695 0.161 0.001 0.936 <0.001 0.002 <0.001 0.283 <0.001 0.203 <0.001 <0.001 <0.001 <0.001 <0.001 0.980 (0.951 to 1.010) 1.674 (0.930 to 3.012) 1.102 (0.715 to 1.699) 1.386 (0.876 to 2.192) 1.396 (0.883 to 2.207) 1.075 (0.694 to 1.644) 1.904 (1.224 to 2.963) 0.583 (0.311 to 1.093) 1.111 ( 0.677 to 2,084) 1,068 (0.603 to 1,889) 1,085 (0.948 to 1,241) 1,296 (1,063 to 1,580) 1,045 (0,888 to 1,230) 0,995 (0.982 to 1,009) 2,197 (1,431 to 3,371) 0.185 0.086 0.688 0.163 0.153 0.746 0.004 0.092 0.617 0.822 0.236 0.010 0.596 0.494 <0.001

1 Yes; 2 A unit in the natural logarithmic scale for the variables troponin T, NT-proBNP, CRP and GDF-15, having been transformed into their natural logarithms before analysis; 3 Variation of 1 ml / min.

1-year mortality according to GDF-15 levels in / at

Baseline 24 hours 48 hours 72 hours GDF-15 <1200 ng / l 1 (0.8) 3 (2.5) 3 (2.7) 5 (4.5) GDF-15 1200-1800 ng / l 3 (2.4) 0 (0) 5 (4.3) 2 (1.7) GDF-15> 1800 ng / l 19 (12.7) 20 (12.1) 15 (8.7) 16 (9 , 5) P value * <0.001 <0.001 0.037 0.051

In a cohort of 399 patients, GDF-15 levels were determined upon admission (baseline) and at 24, 48 and 72 hours. At each time the patients were stratified into tertiles according to their GDF-15 levels. The number of deaths (%) at 1 year is recorded per tertile of GDF-15. * Cochran-Armitage trend test.

GDF-15 levels and risk of recurrent myocardial infarction in the SCA-NSTE

The level of GDF-15 upon admission was also remarkably related to the risk of the outcome of death or recurrent myocardial infarction at 30 days. With increasing tertiles of GDF-15, the risk of death or 10 myocardial infarction at 30 days was 5.0, 6.9 and 10.8%, respectively (P <0.001). This relationship was mainly derived from the association of GDF-15 with mortality. Although the risk of another myocardial infarction within 30 days after the incident was significantly related to increasing tertiles of GDF-15, significantly with rates of 4.8, 5.6 and 7.2%, respectively (P = 0.048) , there was no independent relationship between the GDF-15 and the recurrent myocardial infarction rate at 30 days in the regression analysis

15 multiple logistics.

Combination of GDF-15 with prognostic markers in the SCA-NSTE

Multiple logistic regression analyzes demonstrated that GDF-15 and NT-proBNP were the only important biomarkers for an independent prognosis of mortality. Therefore, tertiles of these markers were combined and used as the main independent clinical predictor of mortality, before myocardial infarction. The results reveal that a combination of increasing tertiles of GDF-15 and NTproBNP levels provided additional prognostic information and identified strata of patients with a 1-year mortality rate between 0.3 and 13.7% in patients without previous history of myocardial infarction and between 4.7 and 23.6% in patients with a previous history of myocardial infarction (Figure 4). These relationships were further modified by the patient's age, with a relative variation of 4.4% in the risk of mortality for each year.

25 years old (table 4).

Example 3: Usefulness of GDF-15 for the prognosis of patients with chronic heart failure (this example is not part of the present invention)

The relationship of GDF-15 levels with clinical and biochemical parameters at baseline and survival in a cohort of 235 patients with FCC enrolled in four European centers, in 30 Athens (Greece, n = 51), was initially investigated. , London (United Kingdom, N = 89) and Wroclaw / Zabre (Poland, n = 95) (derivation cohort). In the validation cohort we prospectively evaluate the hypothesis of the derivation cohort principle, that is, that increasing levels of circulating GDF-15 provide independent prognostic information in patients with FCC. The validation cohort included 220 patients with FCC recruited in Verona (Italy). All patients participated in projects designed to investigate new neurohormonal and inflammatory biomarkers in the prognosis of chronic heart failure and gave their written consent. In all patients, the diagnosis of FCC was based on symptoms and clinical signs and on tests of dilatation or systolic dysfunction of the left ventricle by radionuclide or invasive ventriculography, or by echocardiography. All patients had a minimum history of six months of FCC and were stabilized by medication, at least four weeks before the study. Patients with myocardial infarction in the previous 12 weeks, disease

40 known inflammatory or malignant or creatinine levels> 400 µmol / l were excluded. The institutional ethical committees of all the participating sites in the study approved the protocol.

Patient follow-up was done through outpatient assessments and telephone contact. Survival status was inspected on May 25, 2005 in the referral cohort and on May 31, 2006 in the validation cohort. No patients were left without follow-up. The main criterion of the study was mortality due to

45 all causes. Information on the cause of death was also available in the validation cohort. At the time of the inspection, nine patients underwent a heart transplant were registered.

After a rest time ≥10 minutes in a semi-existent position, venous blood samples were taken to determine the GDF-15 and other parameters. GDF-15 concentrations were determined by an immunoradiometric assay (IRMA), using a goat anti-human GDF-15 polyclonal IgG antibody, purified by GDF-15 affinity chromatography, from R&D Systems (AF957), as described above. recently described. All measurements of GDF-15 were performed at the Medical School of Hannover by researchers who did not know the characteristics of the patients or their results. NT-proBNP levels were determined by a chemiluminescence immunoassay (ELICIA, Roche Diagnostics). The measurements of creatinine, uric acid and hemoglobin were made in the centers participating in the study. Creatinine clearance was calculated according to

55 the equation of Cockcroft and Gault.

or absolute numbers and percentages, as appropriate. The Kolmogorov-Smimov test was used for the normal distribution of continuous variables. Continuous variables were compared using the Mann-Withney test or Student's t-test for unpaired data. Comparisons between patient strata were carried out 5 using the Kruskal-Wallis test or ANOVA. The proportions were compared using the chi-square test. Multiple regression analyzes were applied to identify associated factors independently at GDF-15 levels. Univariate and multivariate Cox proportional hazards analysis were used to assess forecast associations. Kaplan-Meier graphs were used to illustrate the sequence of episodes during the follow-up of GDF-15 levels and the statistical assessment was performed by Cox regression analysis. For

10 comparing the prognostic values of GDF-15, creatinine, uric acid, hemoglobin and NT-proBNP, receiver operating characteristic curves (ROC) were generated and the areas under the curves (ABC) were calculated. All data analyzes were performed with StatView 5.0.1 or MedCalc 8.2.0.3 statistical programs (ROC analysis).

The referral cohort consisted of 235 patients with a median age of 66 years (25th and 75th percentiles, 56

- 72). The clinical and biochemical characteristics of the patients are summarized in Table 6. The median level

15 of GDF-15 was 2240 (1232-4010) ng / l. Approximately three quarters of the patients (75.3%) had GDF-15 levels greater than 1200 ng / l, the upper normal limit (LNS) in apparently healthy elderly subjects. GDF-15 levels were closely related to the NYHA functional class: NYHA I: 850 (646-1747) ng / l, NYHA II: 1621 (1025-2563) ng / l, NYHA III: 2510 (1385-3686) ng / l, NYHA IV: 4869 (2722-8807) ng / l (P <0.001 for the trend). Since the LNS corresponded to the lower quartile limit of the shunt cohort, patients

20 were stratified into quartiles (cutoff limits of 1200, 2200 and 4000 ng / l) for the other analyzes. Of the 235 patients in the bypass cohort 68 (28.9%) died during follow-up. Mortality rates at 12 and 18 months were 15.0% (95% CI, 10.3-19.7) and 22.9% (95% CI, 16.5-29.3), respectively. The risk of death during follow-up increased markedly with increasing quartiles of GDF-15 (Figure 5A). Mortality rates were 2.0, 8.7, 10.5 and 37.5% at 12 months and 2.0, 23.8, 19.1 and 45.6% at 18 months in the

25 respective quartiles (P <0.001).

The validation cohort consisted of 220 patients with a median age of 63 (58-69) years. Patients in the validation cohort were less symptomatic (NYHA class), had lower NT-proBNP levels and were treated more often with ACE inhibitors or angiotensin receptor blockers, β blockers or diuretics, but were their age , sex distribution, body mass index, etiology of heart failure, LVEF, renal function and their uric acid and hemoglobin levels were comparable with the shunt cohort (Table 6). The median level of GDF-15 was 1465 (1004 - 2194) ng / l, significantly lower than that of the shunt cohort (Table 6); In the validation cohort, 65% of the patients had GDF-15 levels higher than the LNS. GDF-15 levels were closely related to the NYHA functional class in the validation cohort: NYHA I: 1106 (849 - 1682) ng / l, NYHA II: 1294 (1007 - 1945) ng / l, NYHA III: 1903 (1122-2650) ng / l, NYHA IV: 4147 (1971

35 5905) ng / l (P <0.001 for the trend).

Table 6: Patient characteristics

 Referral study Validation study Value P (n = 235) (n = 220)

Clinical features Age [years] 66 (56-72) 63 (58-69) 0.161 Male sex 215 (91.5) 197 (89.5) 0.479 BMI [kg / m2] at 25.4 (23.0 - 29.3) 26.1 (23.7 - 28.7) 0.444 Ischemic Etiology 168 (71.5) 140 (63.6) 0.085 Severity of heart failure and biomarkers NYHA class 2.7 ± 0.8 2.2 ± 0.7 <0.001 I 12 (5.1) 31 (14.1) II 84 (35.7) 111 (50.5) III 96 (40.9) 72 (32.7) IV 43 (18.3) 6 (2.7) LVEF [%] 30 (24-40) 33 (27-38) 0.082 Creatinine [µmol / l] 102 (84-137) 99 (90-115) 0.430 Creatinine removal [ml / min] at 71.5 (45.6 - 90.8) 68.4 (53.8 - 88.8) 0.638 Hemoglobin [g / dl] c 13.8 (12.3-14.9) 14.0 (13.1-14.8) 0.108 NT-proBNP [ng / l] 1340 (434-3740) 521 (209-1070) <0.001 GDF-15 [ng / l] 2240 (1232 - 4010) 1465 (1004 - 2194) <0.001 Medication at baseline <0.001 ACE inhibitor or BRA 183 (77.9) 213 (96.8) 0.004 Blocker β 112 (47.7) 146 (66.4) <0.001 Diuretic 167 (71.1) 194 (88.2) <0.001 Spironolactone 58 (24.7) 50 (22.7) 0.703 Uric acid [µmol / l] b 422 (345-529) 390 (330-450) 0.392

Values are n (%), median (25th and 75th percentiles) or mean ± SD. aData of 216 patients in the cohort of Of the 220 patients in the validation cohort 49 (22.3%) died during follow-up. Rates of

5 mortality at 12, 18 and 60 months were 4.6% (95% CI, 1.9 - 7.3), 6.9% (95% CI, 3.6-10.2) and 18, 6% (95% CI, 13.3-23.9), respectively. When using the cut-off limits established in the bypass cohort, the highest levels of GDF-15 were closely related to all-cause mortality (Figure 5B). Mortality rates were 3.5, 3.8, 8.0 and 6.3% at 12 months, 5.9, 3.8, 10.7 and 26.3% at 18 months and 10.7, 16.2, 27.2 and 53.1% at 60 months, in patients with GDF-15 levels ≤ 1200 ng / l, between 1201 and 2200 ng / l, between 2201 and

10 4000 ng / l and> 4000 ng / l, respectively (P <0.001). Both patients who died from progressive heart failure and those who died suddenly had clearly higher levels of GDF-15 compared to survivors (Table 7).

Table 7: Multivariate analysis of all-cause mortality by Cox regression

Shunt Cohort Validation Cohort

Characteristics TR (95% CI) P Value TR (95% CI) P Value Age (every 10 years) 1,158 (0,906-1,479) 0,241 1,190 (0,776-1,826) 0.425 Male 3,491 (0,787-15.49) 0,100 1,018 (0,369-2,811) 0.973 Ischemic etiology 1,101 (0.627-1.931) 0.738 1.009 (0.525-1.939) 0.979 LVEF (by 10% decrease) 1,609 (1,241-2,086) <0.001 1,843 (1,177-2,885) 0.008 NYHA class (per class) 1,296 (0,889-1,891) 0.177 1.039 (0.640-1.688) 0.877 Creatinine Ln 2,329 (0.939-5.781) 0.068 0.523 (0.102-2.684) 0.437 Ln of uric acid 0.903 (0.398-2.047) 0.806 0.817 (0.239-2.787) 0.747 Hb (per 1 g / dl decrease) 1,094 (0.939-1.273) 0.248 0.952 (0.758-1.197) 0.676 NT-proBNP L 1,076 (0.836-1.386) 0.569 1.153 (0.824-1.613) 0.407 Ln of GDF-15 2,156 (1,307-3,566) 0.003 2,888 (1,345-6,200) 0.007

15 Risk rates (TR) with 95% confidence intervals (CI) and P values are indicated. Creatinine, uric acid, NTproBNP and GDF-15 were not normally distributed and therefore transformed into their natural logarithms before analysis. The risk rates refer to an increase of one unit of the scale of natural logarithms in these variables. Hb means hemoglobin.

20 Several clinical features and biochemical parameters are indicative of a poor prognosis in patients with CHF and the univariate analysis by Cox regression confirmed the usefulness of several of these accredited markers in both patient cohorts (Figure 6, A and B). Advanced age, a high NYHA class, a reduced LVEF, low creatinine clearance and elevated creatinine levels and NT-proBNP were associated with an increased risk of death in both cohorts. High uric acid levels, lower body mass index and low

25 hemoglobin concentrations predicted all-cause mortality only in the shunt cohort. Corresponding to the data represented in Figure 5, the highest levels of GDF-15 were related to a significant increase in mortality in both cohorts. Cox regression multivariate analysis revealed that GDF-15 and LVEF were the only independent predictors of all-cause mortality, both in the shunt cohort and in the validation cohort (Table 7). The results did not vary taking the GDF-15 as

30 variable continuous. In the combined material, GDF-15 (P <0.001) together with LVEF (P <0.001) and age (P = 0.003) appeared as independent predictors. By including in the model the body mass index and creatinine elimination (data obtained from 408 patients), GDF-15 (P <0.001) together with LVEF (P <0.001) and creatinine elimination (P = 0.045 ) independently predicted mortality from all causes. ROC curve analyzes further revealed that GDF-15 is a strong biochemical indicator of mortality, with a

35 area under the curve (ABC) of 0.830 (95% CI, 0.783-0.870), which did not differ significantly (P = 0.523) from the ABC of the NT-proBNP (0.852; 95% CI, 0.806-0.890), but was significantly much higher (P <0.001) than the creatinine ABCs (0.646; 95% CI, 0.589-0.700), uric acid (0.657; 95% CI, 0.598-0.708) and hemoglobin (0.655; 95% CI, 0.598-0.708). The best level of GDF-15 to predict mortality at 18 months in the combined patient population was 2729 ng / l (sensitivity, 81.1%; specificity, 71.2%; probability ratio

40 positive, 2.81). Additional analyzes by ROC curves indicated that 2729 ng / l was the best cut-off value to predict mortality at 12 months (data not shown).

Example 4: Usefulness of GDF-15 for prognosis in patients with pulmonary embolism (this example is not part of the present invention).

The amount of GDF-15 in serum samples from a cohort of 123 patients with embolism has been determined

Acute pulmonary, as described in the previous examples. Patients with acute pulmonary embolism had significantly elevated median GDF-15 values (2,196 pg / ml; 25th-75th percentiles: 1,333 to 3,457 pg / ml) compared to healthy individuals (p <0.001). Specifically, 82% (n = 101) had GDF-15 levels greater than the upper normal limit of 1,200 pg / ml. Patients who developed serious complications (needed intubation, catecholamine administration or cardiopulmonary resuscitation) or who died within 30 days

50 at sampling (n = 17) had serum levels of up to 6,039 pg / ml in the median and 25th-75th percentiles of 2,778 to 19,722 pg / ml, while uncomplicated patients had serum levels of 2,036 pg / ml in the median and 25th-75th percentiles of 1,279 pg / ml to 3,176 pg / ml (p <0.001). The evaluation of the data showed that a level of GDF-15 greater than 4,600 pg / ml indicated a 10-fold risk of developing complications or death (that is, the risk increased from 5.0% to 52.2%; p <0.001); see figure 7 (A). At six months the risk of death had increased to 7.7 times (n = 22); see figure 7 (B).

Claims (9)

1. Method to find out if a subject affected by an acute coronary syndrome needs a cardiac intervention, which consists of a) determine the amount of growth differentiation factor 15 (GDF-15) in a sample of blood, serum or plasma from a subject in need of cardiac intervention and b) compare the amount of GDF-15 found in step a) with a reference amount of 1200 pg / ml, so that an amount of GDF-15 greater than 1200 pg / ml indicates that a subject is susceptible to cardiac therapy and said cardiac intervention is chosen from the group consisting of percutaneous coronary angioplasty, percutaneous transluminal coronary balloon angioplasty, laser angioplasty and coronary stent implantation or bypass implantation.

2.

The method of claim 1, further comprising determining the amount of a cardiac troponin in said subject sample and comparing the amount of cardiac troponin with a reference amount.

3.

The method of claim 2, wherein said cardiac troponin is troponin T.

Four.

The method according to one of claims 1 to 3, further comprising determining the amount of a natriuretic peptide in said subject sample and comparing the amount of the natriuretic peptide with a reference.

5.

The method of claim 4, wherein said natriuretic peptide is NT-proBNP.

6.

The method according to any one of claims 2 to 5, wherein said reference amount is 0.01 ng / ml for cardiac troponin and said cardiac troponin is troponin T.

7.

The method of claim 6, wherein an amount of the cardiac troponin greater than the reference amount indicates that a subject is susceptible to cardiac intervention.

8.

The method according to any one of claims 4 to 7, wherein said reference amount is 1000 pg / ml for the natriuretic peptide and said natriuretic peptide is NT-proBNP.

9.

The method of claim 8, wherein an amount of the natriuretic peptide greater than the reference amount indicates that a subject is susceptible to cardiac intervention.